@Andres4NY@social.ridetrans.it
Post #1058503
2026-04-10 16:42 UTC
Replies (5)
-
@alienghic@timeloop.cafe 2026-04-10 17:16
@Andres4NY @KathyReid Given the way people are using the internet I thought named data networking would be a near option to the current host based networking https://en.wikipedia.org/wiki/Named_data_networking
-
@bms48@mastodon.social 2026-04-10 20:14
@Andres4NY @KathyReid Some of this has in fact already happened, the problem is how you manage and deploy federated networks at scale, but there is a point beyond which that is getting into the realms of "I might want to do this semi-commercially" and "you have to sign an NDA with @bms48 's company". MPLS is a thing. Segment routing is only just becoming a thing. How end-users connect to the network is always a function of the service provider arena: 464 XLAT, PPPoE to name two modern things
-
@bms48@mastodon.social 2026-04-10 20:15
@Andres4NY @KathyReid The sad truth is that some approaches to IPv7 or IPv8 or wotnot is that they require an incremental deployment story to be of any use, and that acts as a filter for many technologies, like TRILL or BGPSec. It's like the Daleks before Sylvester McCoy as the Doctor: it's no use just taking over the downstairs part of the world, mate.
-
@bms48@mastodon.social 2026-04-10 20:17
@Andres4NY @KathyReid Michael Padlipsky's book "The elements of networking style", from his time co-designing the forerunner of the original Internet, the ARPANET, is very illuminating here. It's at the Internet Archive, and the clankers (GenAI, LLMs) haven't trained on it because of the protective LCP PDF DRM. https://archive.org/details/elementsofnetwor00padl
-
@bms48@mastodon.social 2026-04-10 20:17
@Andres4NY @KathyReid Disclaimer: this stuff is directly in the firing line of my actual PhD dissertation